Qwen 3 Coder 480B A35B: benchmark results
Built for repository-scale coding agents that inspect code and call tools. Uses a dedicated function-call format and answers without separate thinking blocks. Provider: Alibaba. Released 2025-07-22. Access: Open.
Aggregate Rating 1665 ± 11, rank #286 of 1635 rated models, from 151 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Multi-SWE-bench flash (MopenHands) | 25.33 | Resolved rate (%) | 100 |
| Multi-SWE-bench mini (MopenHands) | 25.75 | Resolved rate (%) | 100 |
| R2ABench (Mini-SWE-Agent) - GED Accuracy | 51.76 | Normalized graph-edit-distance accuracy (0-100) between the | 100 |
| R2ABench (OpenHands) - Edge F1 | 0.18 | Edge F1 (0-1): F1 of generated dependency and data-flow rela | 100 |
| R2ABench (OpenHands) - GED Accuracy | 52.82 | Normalized graph-edit-distance accuracy (0-100) between the | 100 |
| R2ABench - GED Accuracy | 47.85 | Normalized graph-edit-distance accuracy (0-100) between the | 100 |
| SWE-bench Live Lite (OpenHands) | 24.67 | Resolved rate (%) | 100 |
| TuRTLe - Icarus EM | 38.88 | Average Score (%) | 95.8 |
| TuRTLe - Verilator EM | 38.88 | Average Score (%) | 95.8 |
| BesiegeField - Car - Iterative Editing | 15.24 | Mean distance (m) | 85.7 |
| BesiegeField - Catapult - Iterative Editing | 5.1 | Mean distance (m) | 85.7 |
| TuRTLe - Verilator Syntax | 92.31 | Average Score (%) | 84.1 |
Interactive version: theaggregate.ai/model?slug=qwen-3-coder-480b-a35b · How It Works · Data refreshed daily, snapshot 2026-11-05.